That theory has been around the block but it is in fashion again. In 2009, David Kessler’s book, “The End of Overeating” put forward the thesis that food in contemporary American food has been deliberately engineered–by adding fat, sugar and salt–to exploit our neurochemistry and hijack our free will.
More recently, one of the luminaries of the Paleo movement, Stephan Guyenet, has formulated his own version of this theory, in a compelling series on his Whole Health Source blog, arguing that “food reward” is a main driver of obesity. His prescription: eat a bland diet. Guyenet’s talk about this at the Ancestral Health Symposium last month is the buzz of the paleosphere.
But I think the theory is wrong, for the simple reason that it too blindly takes correlation for causation. And in doing so, it gets the causal direction mostly wrong. We don’t get fat because food has become too tasty. Rather, to a large extent, it is the metabolism and dietary habits of the obese that make food taste too good to resist, leading to insatiable appetites. And the good news is that we are not consigned to blandness. If we eat and exercise sensibly, we can eat flavorful, delicious foods and enjoy life, without packing on the pounds.
Brain chemistry. Stephan Guyenet’s series on food reward, like Kessler’s book, pins the blame for obesity largely on the increased availability of more palatable “high reward” food.
According to USDA data, Americans today eat an astonishing 425 more calories per day than they did in 1970. That is the reason for the obesity epidemic, plain and simple. However, that fact doesn’t tell us why we’re eating more calories, so its usefulness is limited. The increase in calorie intake has come primarily from refined carbohydrate, but even that doesn’t get us very far, because why did we decide to eat more refined carbohydrate? Probably because of the systematic efforts of commercial food manufacturers to increase the palatability/reward value and availability of processed food. In the last four decades, the US has become saturated with hyperpalatable/rewarding commercial and restaurant foods including fast food, soda, french fries, chips, candy and other industrial products. I’ve seen people claim that they ate these things just as much in the 1960s and 70s, but the USDA and National Restaurant Association data show otherwise. The qualitative changes in the US diet have been swift and profound… (A Roadmap to Obesity, August 25, 2011)
But what is it about food that makes it rewarding or not? Guyenet suggests that food reward relates to opioid and dopamine signaling:
Feeling satisfied after eating something is not reward. If you keep eating a starch food beyond what’s appropriate, that is probably because it has too much reward/hedonic value for you. Opioid signaling, implicated in hedonic processing, shuts off satiation signals in the brain and may also increase the setpoint. Dopamine signaling, implicated in reward, can strongly influence food intake and also seems to be able to increase the set point.
In “The End of Overeating”, Kessler also emphasizes the way that “hyperpalatable” foods stimulate dopamine, opioids and other reward neurotransmitters. To be fair, both Guyenet and Kessler acknowledge that food reward is not the only explanation for obesity. They acknowledge the role of genetics, exercise and other factors. But for both of them too-tasty food is the leading culprit.
The relativity of taste. But is it really that simple? Are some foods inherently and invariably rewarding? Do our taste buds and noses directly respond to tasty foods or foods high in fat, sugar or salt by stimulating the secretion of dopamine and opioids in the brain, turning us into addicts? Somehow, it must be more complex than that.
Seth Roberts has postulated a different explanation, in which learning plays a role. His Shangri-La Diet was derived from observations that tasty foods lead to weight gain only after repeatedly encountering the flavor and the calories together. Roberts calls this process “flavor-calorie association”. It’s a Pavlovian conditioning process: the more habitual the association, the greater the obesogenic potential of the food or beverage. So his diet involves regular doses of “flavorless calories” in the form of bland oils, sugars or proteins. Alternative strategies include consuming foods with unfamiliar flavors or “crazy spices”, or flavored noncaloric beverages like herb teas. (For more on flavor-calorie association, see my post on Flavor Control Diets).
Some foods and flavors may be naturally appealing to infants and children, but there is strong evidence that food preferences vary considerably among individuals and cultures. “One man’s food is another man’s poison”. Roberts describes the interesting story of a Gaku Homma, Japanese cookbook author whose initial impression of Coke was that it tasted “like medicine” and was repulsed by it. Similarly, Westerners are often repulsed by Asian fermented foods, or delicacies like dog or snake. Certain cultures find insects and grubs to be delectable, but most of us would probably pass, even knowing that such foods represent a nutritious source of calories. There are many unfamiliar foods, rich in fat, sugar, salt or flavor, that the average fan of potato chips and ice cream would reject, even if hungry. For an interesting discussion of the cultural relativity of food acceptance and rejection of unfamiliar or novel foods, see the article by John Prescott in the Encyclopedia of Food and Culture.
Likewise, over time we can learn to like flavors and tastes that were once unappealing. Roberts cites experiments where rats that do not like the taste of saccharine, grow to like it when they are intravenously fed glucose. There are many “acquired tastes” that we come to like only after repeated exposures. Our palates are changeable.
If you think that food aromas are naturally or inherently appetizing, rather than relative, ask yourself: Why do we respond to food odors differently than other evocative and pleasant odors – flowers and plants, soil, sea, even pleasant or sensual human scents? Smelling a rose does not make you hungry. Could we be conditioned to salivate and secrete insulin in response to the smell of a rose if we always sniffed a rose before gulping down a sweet drink? I think so. Pleasant aromas or tastes don’t necessarily generate a drive to eat. The association between sensation and the drive to eat must be learned.
In fact, both Guyenet and Kessler appear to acknowledge the relativity of taste at certain points in their accounts. For example, Guyenet notes that taste preferences towards beer or vegetables change as we transition from childhood to adulthood. It is instructive that Guyenet defines “food reward” in a surprisingly broad way:
Food reward is the process by which eating specific foods reinforces behaviors that favor the acquisition and consumption of the food in question. You could also call rewarding food “reinforcing” or “habit-forming”, although not necessarily in an addictive sense. Food reward is a perfectly normal and healthy part of life, although I believe it can be harmful if it exceeds the bounds of what we’re adapted to. Food reward is essential for survival in a natural environment, because it teaches you what to eat and how to get it through a trial-and-error process. (Food reward, May 26, 2011)
In this definition of reward, Guyenet seems to move away from the idea that “reward” is an inherent property of food (i.e. fat, sugar, salt) in triggering opioids and dopamine, but rather is a result of conditioning. Sounding very similar to Roberts, Guyenet notes that
Researchers have demonstrated in rodents and humans that pairing a flavor with a source of calories makes us gradually enjoy the flavor more, whether or not it remains paired to calories afterward. That’s called a “conditioned flavor preference”, and it’s a simple demonstration of food reward in action. The brain senses the ingested calories and assigns a positive reward value to the cues (flavor, location, etc.) associated with the calories, after which we’ll be more likely to eat something that contains the preferred flavor.
As another example, rats prefer to hang around a place where they have repeatedly received rewarding food. Have you ever seen a child run after an ice cream truck? After a certain time, our motivation to obtain a food that we perceive as rewarding increases, and so does our consumption of it. Rats accustomed to eating human junk food will endure foot shocks and extreme temperatures to obtain it, even when much healthier unprocessed rodent chow is freely available
Put another way:
It doesn’t matter whether or not you like the Little Debbie cake once it’s in your mouth. It doesn’t matter how you feel afterward. The only thing that matters is whether or not you’ll buy another one tomorrow. That’s food reward.
Kessler also acknowledges the role of Pavlovian conditioning in appetite, recognizing that not just flavors, but any cues can serve as reinforcers. In Chapter 10 of his book, he cites Pavlov’s success in training dogs to salivate in response to the ringing of a bell, even after it is no longer accompanied by food. Kessler discusses Kent Berridge’s related concept of “incentive salience” :
Simply put, incentive salience is the desire, activated by cues, for something that predicts reward. It’s a learned association — we learn to want a food or some other substance we once liked…Cue-induced wanting, said Berridge is “triggered by the sight of a cookie or someone lighting up a cigarette nearby or clinking the ice cubes in the glass of alcohol…Those kinds of cues have the power to evoke the desire to take that thing again.” Experience imbues the cue with incentive salience. Positive emotions become embedded in cues, which then develop a force of their own. (The End of Overeating, Ch. 10)
So the door has been opened here to the idea that “reward” is not an inherent property of food but rather the consequence of a conditioned association or “pairing” between the calories in the food and a sensible signal or cue. The cue could be a flavor, but it could just as well be a visual or auditory cue, a familiar location or a social context. In this understanding, “reward” not an inherent property of the food, but is rather a learned response to perceptual cues associated with the food. These cues need only at some point to have become associated with a conditioned expectation of caloric value.
But now we come to the internal contradiction in the Kessler-Guyenet theory of food reward: Is “reward” objective and invariant — or relative, subjective and variable? It cannot be both. To say that reward is “relative” means that it varies markedly among individuals and across cultures, but that does not make it any less real. It can be a powerfully motivating force, driving the obese to overeat and even binge to unhealthful extremes. But to acknowledge the subjectivity and relativity of food reward is at odds with the idea that there is such thing as inherently “hyperpalatable” food that is irresistibly obesogenic in and of itself.
An alternative explanation: impaired metabolism. I’m not denying here that people crave or get addicted to foods like potato chips, cookies and ice cream. No doubt these people find the flavors salient and compelling, even to the point of addiction. But it’s not the flavor that causes the behavior in the first place. The flavor only becomes a strong cue under certain conditions. That’s obvious from the simple fact that many people, eating the very same foods, do not find them to be addictive. While I occasionally enjoy a cookie or some ice cream, I actually find it repulsive to eat more than a modest amount. While I used to like soda, I now experience Coke as sickly sweet. And I think I’m not alone in that reaction.
A more likely explanation is that food addicts have altered, perhaps even damaged, their metabolisms. They are most likely insulin-resistant and leptin-resistant as a result of many possible factors, including obesity, stress and inflammation of their insulin receptors and glucose transporters. Guyenet aptly describes how inflammation and lipotoxicity damage the hypothalamus:
There’s an additional factor that I’ve come to believe may be an “elephant in the room” when it comes to insulin/leptin resistance and chronic inflammation, and that is, ironically, energy excess. Glucose and fatty acids, the body’s main two fuels, are toxic when present in the bloodstream in excess. When someone eats too many calories, his body has to deal with the excess. The healthiest way of doing this is actually to shunt the excess energy into fat tissue where it is inert. If the fat tissue does not have a sufficient affinity for the excess fat, free fatty acid levels in the circulation may rise, and tissues and cells may accumulate fat and fat metabolites…if fat mass increases enough, fat cells become insulin resistant, release more fatty acids into the circulation and fail to clear fatty acids from the circulation after a mixed meal. Essentially, fat tissue loses its formerly high affinity for excess fat, and these undesirable fat metabolites accumulate in lean tissues in a manner reminiscent of lipodystrophy. This contributes to insulin resistance and glucose intolerance by the same mechanism described above, creating an excess of circulating glucose as well, which together with the excess of fatty acids can enhance chronic inflammation, further insulin resistance and damage the insulin-secreting pancreas.
…Therefore, it’s possible that an excess of circulating fatty acids (and perhaps glucose) itself acts to raise the setpoint through the gradual accumulation of fatty acid metabolites and inflammation in the hypothalamus, promoting leptin resistance and creating a “cascading failure” of energy balance regulation, glucose metabolism and inflammatory signaling. This would explain why people in affluent societies have trouble staying lean as they age, as well as why obesity is so difficult to treat. I think this is likely to be a late stage process, occurring after significant body fat accumulation and essentially “cementing” the increase in body fatness. The early stage that causes the initial rise in body fatness probably has more to do with food reward/palatability/availability, although that should remain a factor even after obesity is well established…The basic idea is that in genetically susceptible people, excessive food reward/palatability/availability and inactivity cause overconsumption and an increase in the body fat setpoint, followed by the eventual accumulation of fat metabolites and inflammation in the hypothalamus, which exacerbate the problem and make it more difficult to treat. Other factors, such as micronutrients, gut flora, fiber, fat quality, polyphenols, sleep and stress, may also play a role. I think this is a reasonable working hypothesis of why obesity has increased so rapidly in the last 30 years, and is so difficult to treat once established. I believe these ideas are broadly consistent with the research and opinions of senior obesity researchers I respect. (Roadmap to Obesity, August 25, 2011)
Cause and effect. Where I believe Guyenet goes wrong in the above passage is in postulating that “the early stage” of obesity is driven by primarily by food reward causing overconsumption, which then leads to obesity and leptin resistance. I think the cause and effect relationship is reversed. Certainly “normal” reward is part of a healthy appetite, but that doesn’t lead to obesity. It is the leptin resistance of obesity that sets one up for food reward to become pathogenic. According to Robert Lustig, the normal satiating effect of insulin within the brain (CNS) becomes impaired in those with insulin and leptin resistance:
Although CNS insulin levels tend to reflect serum insulin levels, the relationship breaks down in obesity states. In obesity, there is proportionally less CNS insulin; the expression of the CNS insulin transporter is decreased in several obesity models. This paucity of insulin available for satiety signaling may represent a form of CNS insulin resistance. (Lustig, Fast Food, Central Nervous System Insulin Resistance, and Obesity)
Put simply, it takes more time and larger quantities food or beverage for insulin-resistant individuals to become sated, because the appetite-suppressing or “shut off” effect of insulin in the hypothalamus is impaired. Added to this is the fact that obese, insulin-resistant individuals typically have a grossly amplified preprandial insulin response, which means that blood glucose is more easily stoked by mere appetite cues like the sight, aroma, or even thought of food. Frequent, regular and familiar eating of these reward foods further strengthens the reinforcement. Even stress can trigger this hunger cycle. This leads to a very strong drive to start eating and great difficulty in shutting off the eating. And with more overconsumption of food and the resultant obesity, a vicious cycle sets in, leading to heightened insulin-resistance and leptin-resistance. Once this vicious cycle begins, the psychological component of food cravings and addictions is enhanced. The association between flavor cues — or any cues — and consumption of the food is strengthened. And the food becomes more and more palatable, even “hyperpalatable”. But it is the impaired metabolism of obesity and the reinforcing eating patterns that make these foods hyperpalatable to the individual, not the other way around.
Foods are not inherently hyper-rewarding. Rather, an impaired metabolism, combined with reinforcing eating patterns lead to food becoming hyper-rewarding.
I can anticipate the following objection to my argument: Am I saying that any food can become hyperpalatable or addictive? If so, why are foods like french fries, bread, cookies, chocolate and ice cream craved more than celery, cucumbers and lamb chops? The answer, I think, is that if you are leptin-resistant and insulin-resistant, these high calorie foods provide a sufficiently rapid “bolus” injection of calories into the bloodstream to overcome any initial preprandial drop in blood glucose, and to spike insulin sufficiently high to overcome the CNS insulin resistance and thus satisfy appetite. However–and this is a key point–not everyone finds junk foods to be irresistable. Most insulin sensitive individuals are readily sated on pizza or dessert. And not everyone even finds these foods to be enjoyable.
Another important factor in the addictiveness of food is the reduction or impairment in dopamine receptors in the brains of the obese, as documented by PET scans. Interestingly, a similar reduction in dopamine receptors is seen in drug addicts and depressed individuals. This could be a result of the overstimulation by raised levels of dopamine from the “bolus” of large meals or binges. resulting in a homeostatic downregulation of receptors. I’ve discussed this in more detail in my post Change your receptors, change your set point. Probably any large amount of calories, even with unfamiliar, less palatable, or weaker flavors would do the same trick. For a short time, there would be frustration due to reduced dopamine signaling — until the brain learned the new flavor-calorie association.
Bland food diet. From his theory of food reward, Guyenet proposes a way out: Eat bland foods. He supports this by citing evidence that pre-industrial cultures such as the Kitavans eat a diet that is quite bland, despite being high in carbohydrates, and thereby they remain lean and healthy. He references studies on the effects of bland food diets in lean and obese humans (by Hashim and Van Italie, and by Michael Cabanac) to support his thesis.
Investigators have known for decades that the cafeteria diet is a highly effective way of producing obesity in rodents, but what was interesting about this particular study from my perspective is that it compared the cafeteria diet to three other commonly used rodent diets: 1) standard, unpurified chow; 2) a purified/refined high-fat diet; 3) a purified/refined low-fat diet designed as a comparator for the high-fat diet. All three of these diets were given as homogeneous pellets, and the textures range from hard and fibrous (chow) to soft and oily like cookie dough (high-fat). The low-fat diet contains a lot of sugar, the high-fat diet contains a modest amount of sugar, and the chow diet contains virtually none. The particular high-fat diet in this paper (45% fat, which is high for a rat) is commonly used to produce obesity in rats, although it’s not always very effective. The 60% fat version is more effective.
Consistent with previous findings, rats on every diet consumed the same number of calories over time… except the cafeteria diet-fed rats, which ate 30% more than any of the other groups. Rats on every diet gained fat compared to the unpurified chow group, but the cafeteria diet group gained much more than any of the others. There was no difference in fat gain between the purified high-fat and low-fat diets.
So in this paper, they compared two refined diets with vastly different carb:fat ratios and different sugar contents, and yet neither equaled the cafeteria diet in its ability to increase food intake and cause fat gain. The fat, starch and sugar content of the cafeteria diet was not able to fully explain its effect on fat gain. However, each diets’ ability to cause fat gain correlated with its respective food reward qualities. Refined diets high in fat or sugar caused fat gain in rats relative to unpurified chow, but were surpassed by a diet containing a combination of fat, sugar, starch, salt, free glutamate (umami), interesting textures and pleasant and invariant aromas.
Guyenet’s interpretation is that the rats ate more of the “cafeteria diet” because it was more palatable, presumably due to the higher fat, starch and sugar content, than the equally calorie dense blander diets. But this is not proven. How do we know it was more “palatable”, if this is a subjective quality? All we know is that more of the cafeteria diet was consumed. We can of course define that as palatability, but that would make the argument circular. The real question is: Do we eat more calories because inherent “palatability” or taste characteristics? Or do foods become perceived as more palatable because of prior food experience, eating patterns and associations that modify neural circuitry and the drive to eat? Palatability appears not to be something inherent in food, but rather something changeable. We do not know what diets or reinforcement schedules the rats were raised on before Cabanac’s conducted his experiments.
From these and other observations, Guyenet concludes:
Some people may be inclined to think “well, if food tastes bad, you eat less of it; so what!” Although that may be true to some extent, I don’t think it can explain the fact that bland diets affect the calorie intake of lean and obese people differently. To me, that implies that highly rewarding food increases the body fat setpoint in susceptible people, and that food with few rewarding properties allows them to return to a lean state. (Food Reward, A Dominant Factor in Obesity, Part II)
In recognizing that bland diets have different effects on the lean and the obese, Guyenet here seems to made a full retreat from asserting the explanatory power of food reward as a primary driver. The relativity of taste here reveals that it must be a consequence, not a determinant, of metabolism and neural conditioning.
What can we do? If you believe that tasty food is inherently addictive, it is reasonable to seek out bland food, and avoid strong flavors, fat, sugar and salt. But is this necessary? Do we have to give up not just “junk foods” like Big Macs, french fries and ice cream — but also more healthful foods that are flavorful, fatty, sweet or salty? What about lamb curry (fatty and flavored), berries and cream (sweet and fatty) or salted steak? I think not. Flavor, fat, salt and even a modest amount of sugar is not the seed of obesity. Rather, it is the effect that foods have on our hormones and receptors that we should think about. To avoid obesity, we should strive to maximize our insulin sensitivity and leptin sensitivity. This can be done by a variety of measures, discussed elsewhere on this blog, including:
- weight loss, particularly abdominal fat
- intermittent fasting
- avoiding inflammatory foods and toxins that impair receptor sensitivity
- supplementing with fish oil, magnesium and vitamin D for receptor health
- avoiding chronic stress, but pursuing intermittent, intense “good” stress, such as: